Generative Perturbation Network for Universal Adversarial Attacks on Brain-Computer Interfaces.

Deep neural networks (DNNs) have successfully classified EEG-based brain-computer interface (BCI) systems. However, recent studies have found that well-designed input samples, known as adversarial examples, can easily fool well-performed deep neural networks model with minor perturbations undetectable by a human. This paper proposes an efficient generative model named generative perturbation network (GPN), which can generate universal adversarial examples with the same architecture for non-targeted and targeted attacks. Furthermore, the proposed model can be efficiently extended to conditionally or simultaneously generate perturbations for various targets and victim models. Our experimental evaluation demonstrates that perturbations generated by the proposed model outperform previous approaches for crafting signal-agnostic perturbations. We demonstrate that the extended network for signal-specific methods also significantly reduces generation time while performing similarly. The transferability across classification networks of the proposed method is superior to the other methods, which shows our perturbations' high level of generality.

Chelate effects in glyme/lithium bis(trifluoromethanesulfonyl)amide solvate ionic liquids. I. Stability of solvate cations and correlation with electrolyte properties.

To develop a basic understanding of a new class of ionic liquids (ILs), "solvate" ILs, the transport properties of binary mixtures of lithium bis(trifluoromethanesulfonyl)amide (Li[TFSA]) and oligoethers (tetraglyme (G4), triglyme (G3), diglyme (G2), and monoglyme (G1)) or tetrahydrofuran (THF) were studied. The self-diffusion coefficient ratio of the solvents and Li(+) ions (Dsol/DLi) was a good metric for evaluating the stability of the complex cations consisting of Li(+) and the solvent(s). When the molar ratio of Li(+) ions and solvent oxygen atoms ([O]/[Li(+)]) was adjusted to 4 or 5, Dsol/DLi always exceeded unity for THF and G1-based mixtures even at the high concentrations, indicating the presence of uncoordinating or highly exchangeable solvents. In contrast, long-lived complex cations were evidenced by a Dsol/DLi ∼ 1 for the longer G3 and G4. The binary mixtures studied were categorized into two different classes of liquids: concentrated solutions and solvate ILs, based on Dsol/DLi. Mixtures with G2 exhibited intermediate behavior and are likely the borderline dividing the two categories. The effect of chelation on the formation of solvate ILs also strongly correlated with electrolyte properties; the solvate ILs showed improved thermal and electrochemical stability. The ionicity (Λimp/ΛNMR) of [Li(glyme or THF)x][TFSA] exhibited a maximum at an [O]/[Li(+)] ratio of 4 or 5.